Using Number::Phone to validate and format - perl

Question

I'm trying to use Number::Phone from CPAN to accomplish 2 tasks:
Validate a Phone Number; and
Format the number in E.164 Notation.
However, I'm unable to figure out how it works. My sample code is:
#!/usr/bin/perl -w
use strict;
use warnings;
use Number::Phone;
foreach my $fnum ( '17888888', '97338888888', '00923455555333', '[phone]' , '38383838') {
my $phone = Number::Phone->new($fnum);
my $norm = "";
eval {
$norm = $phone->format_using('E123'); # or 'Raw'
print "E164 => '$norm'\n";
} or do {
print STDERR "Unable to parse '$fnum'\n";
}
}
Expected output:
E164 => '[phone]'
E164 => '[phone]'
E164 => '[phone]'
E164 => '[phone]'
E164 => '[phone]'
But the results were incorrect. I tried using Number::Phone::Normalize, but still not successful:
#!/usr/bin/perl -w
use strict;
use warnings;
use Number::Phone::Normalize;
my %params = (
'CountryCode'=>'973',
'IntlPrefix' =>'00',
'CountryCodeOut'=>'973',
'IntlPrefixOut' => '+',
);
my $nlz = Number::Phone::Normalize->new( %params );
foreach my $number ('17888888', '97338888888', '00923455555333', '[phone]' , '38383838') {
my $e164 = $nlz->intl( $number );
print "E164 => '$e164'\n";
}
with the same expected output of:
E164 => '[phone]'
E164 => '[phone]'
E164 => '[phone]'
E164 => '[phone]'
E164 => '[phone]'
However, this produced the wrong results too. The snippet Java code below works perfectly, and it's what I'm trying to achieve in Perl.
// Uses libphonenumber: http://code.google.com/p/libphonenumber/
// setenv CLASSPATH .:libphonenumber-8.5.2.jar
// libphonenumber
import com.google.i18n.phonenumbers.PhoneNumberUtil;
import com.google.i18n.phonenumbers.Phonenumber.PhoneNumber;
import com.google.i18n.phonenumbers.NumberParseException;
import com.google.i18n.phonenumbers.PhoneNumberUtil.PhoneNumberFormat;
public class ValidateList {
public static void main(String[] args) {
try {
if (args.length != 1) {
throw new IllegalArgumentException("Invalid number of arguments.");
}
String file = args[0];
PhoneNumberUtil phoneUtil = PhoneNumberUtil.getInstance();
try (java.io.BufferedReader br = new java.io.BufferedReader(new java.io.FileReader(file))) {
String line = null;
while ((line = br.readLine()) != null) {
try {
PhoneNumber phoneNumber = phoneUtil.parse(line, "BH");
boolean isValid = phoneUtil.isValidNumber(phoneNumber);
if (isValid) {
System.out.println( "E164 => " + phoneUtil.format(phoneNumber, PhoneNumberFormat.E164) );
}
else {
System.err.println( "Invalid => " + line);
}
}
catch (NumberParseException e) {
System.err.println("NumberParseException for ("+line+"): " + e.toString());
}
}
}
}
catch (Exception e) {
System.err.println(e);
System.err.println("Usage: java ValidateList <fileNameWithPhoneNumbers>");
}
}
}
% cat input.txt
17888888
97338888888
00923455555333
[phone]
38383838
% javac -cp libphonenumber-8.5.2.jar ValidateList.java
% java -cp .:libphonenumber-8.5.2.jar ValidateList input.txt
E164 => [phone]
E164 => [phone]
E164 => [phone]
E164 => [phone]
E164 => [phone]
Your input is greatly appreciated.

Answer 1

When I run the first example code for the numbers, two of those fail to be parsed:
17888888 - this is obvious, when calling Number::Phone without a country code, this will not be parsed as it's unclear what country this is from
00923455555333 - 923 is, according to a quick google search, the country code for Pakistan. The Wikipedia page for dialing codes in Pakistan shows no 455, leading me to think that this is not a known area code to either Number::Phone or Wikipedia. I suspect it is an invalid number.
So for the first Number: specify which country this is supposed to be from.
If you are certain the other number is correct, you know more about that than the developer of Number::Phone currently and I'm sure he'd be happy to receive your input in the form of a more complete Number::Phone localized package.
The fact that your Java code accepts the (probably) invalid number does not necessarily mean it is more correct, just that it is less picky in what it declares to be a correct number.
Edit:
Asking Phone::Number to parse the input '[phone]' instead of '00923455555333' leads to the desired output.
Looking at the source of Phone::Number:
# ... processing input arguments
$number = "+$number" unless($number =~ /^\+/);
It becomes clear that the 00 is interpreted as '+00' and then rejected as being an invalid number.
View some discussion on that here
It seems to me you will have to handle this yourself.
One way may be to simply replace leading 00 with '+' - preferably only if parsing failed.
The other number can be parsed if you make it clear what country it should belong to.
Perhaps like so:
my $phone = Number::Phone->new($fnum);
unless ($phone){
$phone = Number::Phone->new('BH',$fnum);
if ( !$phone && $fnum =~ s/^00/+/ ){
# You should probably check the discussion I linked.
# There may well be problems with this approach!
$phone = Number::Phone->new($fnum);
}
}